Yang Zhang
Yang Zhang
> > > > > Thank you for your detailed review again!I have fixed some naming problems and now make sure the pytest will work.However,with only TN part I have...
@mzxcpp 1. could you please remove the unrelated files from your PR: 2. Could you fix the LGTM errors? like removing unused imports etc?  After that, i think we...
> @yzhang123, please also see this https://github.com/NVIDIA/NeMo/pull/4610/files yes i saw this. do you want us to use this for english too? it is a heuristic and will break in some...
when will this be released?
@DachengLi1 , might be a seed issue, i reran from scratch and it was ok. another fix is to swich to fp16 from bf16 or use full attention bias +...